12340
J . Am. Chem. SOC. 1993,115, 12340-12345
Biosynthesis of Esperamicin AI, an Enediyne Antitumor Antibiotic Kin Sing Lam,' Judith A. Veitch, Jerzy Golik, Bala Krishnan, Steven E. Klohr, Kevin J. Volk, Salvatore Forenza, and Terrence W. Doyle Contribution from the Bristol-Myers Squibb Company, Pharmaceutical Research Institute, Wallindord, Connecticut 06492 Received August 30, 1993"
Abstract: Biosynthetic studies on esperamicin At (esp AI) were carried out by examining the incorporation of singly and doubly W-labeled acetates, L- [methyl-Wlmethionine and N a ~ ~ ~ Sby0 cultures 4, of Actinomadura verrucosospora MU-5019. The acetate incorporation results show that the C15 bicyclic enediyne core of esp AI is derived from head-to-tail condensation of seven acetate units and the uncoupled carbon attached to the trisulfide unit is derived from the C2 of acetate. The L- [methyl-l3C]methionine incorporation result shows that the S-methyl groups of the trisulfide and the thiosugar and the O-methyl groups of the aminosugar, the aromatic chromophore, and the carbamate moiety are derived from L-methionine via S-adenosylmethionine. Using Na234S04as the sole sulfur source in the fermentation and by mass spectrometric analysis, we have demonstrated that all four sulfur atoms in esperamicin AI, (esp Alc) can be derived from N a ~ 3 ~ S 0 4On . the basis of the IT-labeled acetate-enrichment pattern, the enediyne ring moiety of esp A1 may be derived from an octaketide with the loss of the C1 of the end acetate unit. The acetate-enrichment pattern of the enediyne moiety of esp A1 is in good agreement with that of dynemicin A (DNM-A). The two carbons comprising the yne moietiesof esp AI and DNM-Aare derived from separate acetateunits. The corresponding carbons inchromophore A of neocarzinostatin (NCS Chrom A) are derived from the same acetate units. This may suggest that enediyne cores of esp AI and DNM-A are biosynthesized from a common precursor while N C S Chrom A is biosynthesized via a different process.
Esperamicin AI (esp AI, Figure l), an extremely potent antitumor antibiotic, was isolated from cultures ofActinomadura verrucosospora. The isolation and theelucidation of the structure of esp A1 have been reported.' Esp A1 consists of a bicyclic core to which are attached a trisaccharide and a substituted 2-deoxyL-fucose with an aromatic chromophore attached to the sugar 3 position. The CISbicyclic core contains the enediyne, an allylic trisulfide, and a bridgehead enone. Detailed structure-activity studies have established that the interaction of these three functionalities results in a bioreductively activated, highly efficient DNA strand scission.z The mechanism of action of esp A1 and several related compounds, calicheamicin,3 neocar~inostatin,~ dynemicins,S and kedarcidin,6 has been reported. It appears that these compounds are capable of cleaving DNA via direct carbon radical abstraction of deoxyribose hydrogen atoms. The presence of the enediyne function in these compounds is a prerequisite to @Abstractpublished in Aduance ACS Abstracts, December 1, 1993. ( I ) (a) Konishi, M.; Ohkuma, H.; Saitoh, K.-I.; Kawaguchi, H.; Golik, J.; Dubay, G.; Groenewold, G.; Krishnan, B.; Doyle, T. W. J . Antibiot. 1985,38, Kawaguchi, 1605-1609. (b) Golik, J.;Clardy, J.;Dubay,G.;Groenewold,G.; H.; Konishi, M.; Krishnan, B.; Ohkuma, H.; Saitoh, K.-I.; Doyle, T. W. J . Am. Chem. Soc. 1987,109,3461-3462. (c) Golik, J.; Dubay, G.; Groenewold, G.; Kawaguchi, H.; Konishi, M.; Krishnan, B.; Ohkuma, H.; Saitoh, K.-I.; Doyle, T. W. J. Am. Chem. SOC. 1987, 109, 3462-3464. (2) (a) Long, B. H.; Golik, J.; Forenza, S.;Ward, B.; Rehfuss, R.; Dabrowiak, J. C.; Catino, J. J.; Musial, S. T.; Brookshire, K. W.; Doyle, T. W. Proc. Natl. Acad. Sci. U.S.A. 1989,86, 2-6. (b) Sugiura, Y.; Uesawa, Y.; Takahashi, Y.; Kuwahara, J.; Golik, J.; Doyle, T. W. Proc. Narl. Acad. Sci. U.S.A. 1989, 86, 7672-7676. (3) (a) Zein, N.; Poncin, M.; Nilakantin, R.; Ellestad, G. A. Science 1989, 244, 697-699. (b) Zein, N.; Sinha, A. M.; MacGahren, W. J.; Ellestad, G. A. Science 1989, 240, 1198-1201. (4) (a) Kappen, L. S.; Goldberg, I. A. Biochemisrry 1983,22,4872-4878. (b) Myers, A. G. Tetrahedron Lett. 1987, 28, 44934496. ( 5 ) Sugiura, Y.; Shiraki, T.; Konishi, M.; Oki, T. Proc. Narl. Acad. Sci. U.S.A. 1990, 87, 3831-3835. (6) (a) Leet, J. E.; Schroeder, D. R.; Hofstead, S. J.; Golik, J.; Colson, K. L.; Huang, S.; Klohr, S. E.; Doyle, T. W. J . Am. Chem. SOC.1992, 114, 7946-7948. (b) Leet, J. E.; Schroeder, D. R.; Langley, D. R.; Colson, K. L.; Huang, S.; Klohr, S. E.; Lee, M. S.; Golik, J.; Hofstead, S. J.; Doyle, T. W. J . Am. Chem. SOC.1993,115,8432-8443. (c) Zein, N.; Colson, K. L.; Leet, J. E.; Schroeder, D. R.; Solomon, W.; Doyle, T. W.; Casazza, A. M. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 2822-2826.
0002-7863/93/1515-12340$04.00/0
C
H
3
O
CH30
EspetamidnAl
R = CH(CH&
EsperamcinAlb
R = CHzCH3
Espemmian A1,
R = CH3
d
0
NH
I
Figure 1. Structures of esperamicin AI,esperamicin Alb, and esperamicin AI,.
their potent activities. The biosyntheses of the neocarzinostatin chromophore7 (NCS Chrom A) and dynemicin A* (DNM-A) have recently been reported. The C14 dienediyne moiety of NCS Chrom A is proposed to be derived from degradation of C18 oleate via the oleate-crepenynate pathway rather than by de novo synthesis from acetate.' Tokiwa et a1.* have proposed that DNM-A is derived from an octaketide precursor with a loss of a two-carbon unit from the carboxylate end to yield the heptaketide intermediate. In this paper, we report the pattern of incorporation of various W-labeled precursors into esp Al. A possible biosynthetic pathway of the C15 enediyne core of esp A1 is (7) Hensens, 0. D.; Giner, J.-L.; Goldberg, I. H. J. Am. Chem. SOC.1989, 111, 3295-3299. ( 8 ) Tokiwa, Y.;Miyoshi-Saitoh, M.; Kobayashi, H.; Sunaga, R.; Konishi, M.; Oki, T.; Iwasski, S. J . Am. Chem. SOC.1992, 114, 41074110.
0 1993 American Chemical Society
Biosynthesis of Esperamicin AI
J. Am. Chem. SOC.,Vol. 115, No. 26, 1993 12341
Table 1. I3C NMR Assignments of [ 1-l3C]-and [2-l3C]Acetate-Labe1edDiacetylesperamicin A1 and
of [ 1,2-13C2]Acetate-Labe.eled
lJCc
Diacetylesperamicin AI carbon 1 2 3 4
5 6 7 8
89 PPm
77.7
99.9 84.2 126.8 123.5 88.9 98.1 72.5
[ 1-l3C]acetate (relative intensity) 4.5
134.4
194.8
2.0
13
136.7 131.8 41.0
14 15
84.3
1 3.8 1 3.5 1 4.7 1
4.7 1
6.0
1
165 45.0
73.6 77.0 76.1 45.2 45.5 80.7 80.5
1 3.3
5.8
72.2 72.1
74.5
3.0
1 1
satellites 42.5,45.1 186 187
79.9 80.2 89.4 90.3 89.7 89.8
3.0
4.2
11 12
148.8
[ 1,2-W2]acetate J1c1-13~2
1
1 4.2 1 4.1 1
1 4.3 1
9 10
[2- W ]acetate (relative intensity)
52.2, 53.6 broad obscured 44.0,45.0 45.4 45.2
presented. We also report the labeling pattern of esp A1 and esp AI, by L- [methyl-Wlmethionine and Na2Y304, respectively. Results and Discussion Incorporation of 1%-Labeled Acetate. The major problem in elucidating the biosynthesis of esp A1 is due to its low production in the fermentation. The estimated titer of esp A1 by the parent strain in the original production medium was about 0.05 Mg/ mL.9 Through extensive media development and strain improvement studies, the titer of esp A1 was increased to 25-30 Mg/mL in the production medium H946 by the new strain A. uerrucososporaMU-5019.9 With the improvement bythismutant strain, feeding experiments can be carried out on a 5-10-L scale to obtain sufficient 13C-enriched esp A1 for N M R studies. Incorporation experiments were carried out by feeding 0.2% [ 1-Wlacetate, [2-13C]acetate, and [ 1,2-I3C2]acetate to the cultures of A. uerrucosospora MU-5019. Twenty-five to thirty milligrams of pure W-enriched esp AI can be isolated from a 10-L fermentation. Esp A I is very soluble in CDCl3 but has limited solubility in CH30H. CDClj was used as the solvent for 13C-enriched esp AI in the initial N M R analysis. The signals of C8, C9, and C10 of the enediyne ring were difficult to quantitate because they were broadened in CDC13. The signal of C1 of the enediyne ring was also buried underneath the CDC13 signal. In order to obtain accurate integration of the signals for the above carbons, 13C-enriched esp A1 was first converted to its diacetyl derivativeand CD30D was used as the solvent for N M R analysis. Sharp signals for all 15 carbons of the enediyne ring of thediacetylesp A, were obtained in the N M R spectrum and are shown in Table I. Figure 2 summarizes the 13C-labelingpattern of esp A1 from the [I3C]acetate supplemented cultures. Upon feeding the culture with [I-Wlacetate, we have demonstrated that C1, C3, C5, C7, C9, C11, and C14 of the enediyne ring were enriched (Figure 2a). Peak intensity enhancements at these seven carbons signals were 2.0-5.8-fold (Table I). Using [2-W]acetate, we have demonstrated that C2, C4, C6, C8, C10, C12, C13, and C15 were enriched (Figure 2b). Peak intensity enhancements at these eight carbon signals were 3.0-6.0-fold (Table I). We have clearly shown that all of the 15 carbons of the enediyne ring portion of esp Al were derived from acetate. In order to determine the connectivity of the carbon units of the enediyne ring, doubly enriched [ 1,2-'3C2]acetate was fed to the esperamicin-producing culture. The incorporation pattern of [ 1,2-I3C2]acetate was confirmed by matching of lJccvalues as shown in Table I and Figure 2c. The l J w coupling constants clearly showed that seven (9) Lam, K. S.; Forenza, S.;Veitch, J. A.; Gustavson, D. R.; Golik, J.; Doyle, T. W. In Microbial Metabolites; Nash, C., Hunter-Cevera,J., Cooper, R., Eveleigh, D. E.,Hamill, R., Eds.; WilliamC. Brown Publishers: Dubuque, IA, 1993; pp 261-274.
CHjCOOH OH CHSOSCNH
OOF-At
0
OH CHIOICNH
OOF.AC
0
0 TSO I
COFAC.
ow3
Figure 2. [I3C]Enrichmentpattern of esperamicin A1 from cultures of A . uerrucosospora MU-5019 supplemented with (a) sodium [ 1-l3C]acetate, (b) sodium [2-I3C]acetate, and (c) sodium [ 1,2-W2]acetate.
pairs of carbons were coupled to each other with C15 being the only carbon that was not coupled. There are only four possible folding patterns for a linear CIS unit to yield the enediyne ring system of the esperamicins as shown in Scheme I. On the basis of the result of the connectivities of the carbons in Figure 2c, folding patterns a and d can be eliminated from consideration. From Table I, C14 and C15 have the highest peak intensity enhancements at 5.8-6.0-fold while C11 and C12 have the least peak intensity enhancements at 2.03.3-fold. It is reasonable to assume that C11 and C12 may be the chain termination unit while C 15 is part of the starter acetate unit. If this is true, the CISchain of the enediyne of esp AI can be derived from an octaketide with the loss of the C1 of the acetate unit (Scheme I). The CISchain can then be folded as in path c, and further reaction (path f) would lead to the formation of the enediyne ring of esp AI. This pathway is therefore favored over the alternate possibility in Scheme I (paths band e). Tokiwa et a1.8 have demonstrated that DNM-A is biosynthesized from
12342 J. Am. Chem. SOC., Vol. 115, No. 26, 1993
Lam et al.
Scheme I. Possible Folding Patterns of the Cl5 Enediyne Ring of Esperamicin A1
/
/
[esperamicin Allb culture condition"
+
+
methionine]"
[esperamicin A I ] (Ccg/mL)b
% inhibition
0 0.02% 0.05% 0.10% 0.20%
21.8 12.7 5.1 0 0
41.7 76.6 100 100
L-Methionine was added to the culture at day 4 of the fermentation. bThe titers of esperamicin AI were determined at day 14 of the fermentation.
Table 11. Effect of Cerulenin and Sodium Oleate on Esperamicin A I Production by A. uerrucosopora MU-5019
control (no addition) 0.5 mM cerulenin 0.5 mM cerulenin + 0.1 % sodium oleate 0.5 mM cerulenin 0.5% sodium oleate 1.O mM cerulenin 1.O mM cerulenin + 0.1% sodium oleate 1.O mM cerulenin 0.5% sodium oleate
Table 111. Effect of L-Methionine on the Production of Esperamicin AI by A. uerrucosospora MU-5019
% inhibition
23.3 8.8 8.2
62.2 64.8
6.7
71.3
0 0
100 100
0
100
a Cerulenin and/or sodium oleate were added to the culture at day 4 of the fermentation. The titers of esperamicin A1 were determined at day 14 of the fermentation.
*
two heptaketide chains, which form the enediyne ring and anthraquinone moiety, respectively. Both the enediyne and anthraquinone moieties are derived from seven head-to-tail coupled acetate units. They further proposed a biosynthetic scheme involving a common octaketide intermediate for the formation of the enediyne ring of the esperamicin/calicheamicin/ dynemicin class of antibiotics. Our [Wlacetate enrichement data of the enediyne ring of esp A1 support the above hypothesis. The labeling patterns of the diyne moieties of esp AI and DNM-A are the same. The two carbons comprising the respective yne moieties are derived from separate acetate units. Hensens et aL7 have proposed that the C I dienediyne ~ chain of N C S Chrom A is derived from degradation of oleate via the oleate-crepenynate pathway for polyacetylenes rather than by de nouo synthesis from acetate. Hensens et al.7 further postulated that the Cl5 enediyne ring of esperamicin/calicheamicin could similarly be derived via the oleate-crepenynate pathway. Our data on esp A1 production from cultures of A. verrucosospora MU-5019 supplemented with cerulenin and sodium oleate are at variance with the above hypothesis. Addition of cerulenin (0.5 and 1.0 mM) to the cultures of A . uerrucosospora MU-5019 before the onset of esperamicin synthesis significantly inhibited esp AI production (Table 11). Cerulenin specifically inhibits the 8-ketoacyl-acyl carrier protein synthase of fatty acid and polyketide biosynthesislo and has no effect on the preformed fatty acid. If esp A1 is derived from the oleate-crepenynate pathway, (10) (a) Omura, S . Bacteriol. Rev. 1976, 40, 681-697. (b) Kitao, C.; Tanaka, H.; Minami, S.; Omura, S . J . Antibior. 1980, 33, 711-716.
adding sodium oleate to the cerulenin supplemented culture should provide the precursor for esp A1 and obviate the repression of esp Al synthesis by cerulenin. We did not observe any improvement in esp Al production when adding sodium oleate (0.1 and 0.5%) to the cerulenin supplemented cultures (Table 11). The [14C]acetate labeling patterns of esp A1 in the cultures with or without cerulenin at the active production phase have demonstrated that esp A l is synthesized de nouo from acetate." Furthermore, the two carbons of the yne moieties of NCS Chrom A are derived from the same acetate units,7 indicating that the biosynthetic pathway of the C14 dienediyne ring structure of NCS Chrom A is different from those of esp Al and DNM-A. Incorporation of ~-[methyPC]Methionine. Most biological methylation in bacteria involves L-methionine, the methyl group of which is activated by S-adenosylation with ATP. Numerous reports12 have shown that L-methionine inhibits the production of antibiotics even though L-methionine is the precursor (methyl donor) of the antibiotics. Table I11 shows the effect of L-methionine on the production of esp AI by A. verrucosospora MU5019. L-Methionine, a t a concentration as low as 0.02%, inhibited the production of esp AI by 41.7%. At 0.1% L-methionine concentration, no esp A, can be detected in the fermentation even though there was no effect on growth of the organism and pH of the fermentation. Experiments with radiolabeled anthramycin showed that reduced yields of anthramycin production when L-methionine was added to the culture of Streptomyces refuineus were due to the interaction of anthramycin with the reactive metabolites produced in the L-methionine-supplementedcultures.1b The thiol of L-methionine or its reactive metabolites generated in the fermentation may react with the trisulfide of esp A1 and convert espAl to its aromatized derivative. Inhibition of DNM-A production was observed when more 0.04% L-methionine was added to the culture of Micromonospora chersina.s The above findings indicated that we cannot add a high initial concentration of L- [methyl-I3C] methionine to the culture of A. uerrucosospora for the esp A1 labeling study. Consequently, ~ - [ m e t h y l - W ] methionine was added to the culture on two different days to yield a final concentration of 0.05%. Even though the production of esp A1 was significantly decreased (56.7%), adequate amounts (6.5 mg) of I3C-enriched esp A1 were obtained for N M R analysis. Figure 3 shows the W-enrichment pattern of esp A1 isolated from the L- [methyl-W]methioninefeeding study. The carbons of seven methyl groups were clearly enriched by 2.6-3.2-fold when ~-[methyPC]methioninewas fed into the culture. We also observed low enrichment of the three carbons of the isopropyl group of the aminosugar. Several examples where formation of an ethyl group from two methyl groups of L-methionine have been reported. These include the hydroxyethyl side chains of (11) Lam, K. S.; Gustavson, D. R.; Veitch, J. A.; Forenza, S . J . Ind. Microbiol. 1993, 12, 99-102.
(12) (a) Demain, A. L.; Newkirk, J. F. J . Biol. Chem. 1962,10,321-325. (b) Demain, A. L.; Newkirk, J. F.; Hendlin, D. J . Bacterid. 1963,85, 339344. (c) Mazumdar, S. K.; Kutzner, H. J. Appl. Microbiol. 1962,10, 157165. (d) Rogers, T. 0.;Birnbaum, J. Antimicrob. Agents Chemother. 1974, 5, 121-132. (e) Gairola, C.; Hurley, L. Eur. J . Appl. Microbiol. 1976, 2, 95-101. ( f ) Uyeda, M.; Demain, A. L. J . Ind. Microbiol. 1988, 3, 57-59.
J . Am. Chem. SOC.,Vol. 115, No. 26, 1993
Biosynthesis of Esperamicin A ,
12343
Table IV. Incorporation of 3SS-Precursorsinto Esperamicins in the Complex (H946) and Defined (DF-15) Mediaa % incorporation of radioactivity in esp A1 and esp AI,
i A
medium H946
medium DF- 15
0.095 0.0 19 0.034
1.9
Na2%04
A
1.2
1.7
3sS-Precursors (1 mCi) were added to the growing cultures of A. uerrucosospora MU-5019 at day 4 and day 2, in medium H946 and medium DF- 15, respectively. The major product of esperamicins produced in medium H946 and medium DF- 15was esp A1 and esp Ale, respectively.
A
CH3
A
precursors L- [35S]methionine L- [3SS]cysteine
0
1
Figure 3. [WIEnrichment pattern of esperamicin AI from culture of A. uerrucosospora MU-5019 supplemented with ~-[methyl-l'C]me-
thionine.
thienamycin13and pectamycin14and the C24 ethyl group of certain plant sterols.15 However, formation of an N-isopropyl group from the methyl of L-methionine has not been reported. The possible incorporation of the methyl of L-methionine into the N-isopropyl group of esp A1 is supported by our results from a study of blocked mutants. Two blocked mutants of esp Al, DG-111-10-6 and DG-108-9-3, were isolated which do not produce any esp AI. They each produce one major product of the fermentation, esp Alb and esp AI, (Figure l ) , respectively,16 which differ from esp A1 only in theN-alkylsubstituent ofthe aminosugar. The presence of these blocked mutants may suggest that the formation of esp AI, (methyl), esp Alb (ethyl), and esp Al (isopropyl) involves the sequential addition of methyl groups at the nitrogen and N-methyl sites of the aminosugar. Similar blocked mutants of the related enediyne antitumor antibiotic calicheamicin have recently been reported.I7 This aspect of the work will be further investigated. Biosynthetic Origin of Sulfur. Esp AI contains four sulfur atoms. It contains a thiomethyl sugar in the trisaccharide moiety and an allylic trisulfide in the bicyclic core. Understanding the biosynthesis of the allylic trisulfide is very important in light of what is known of the mechanism of action of the esperamicins. Studies have shown that the thiolate anion, generated by the reduction of the trisulfide group, undergoes subsequent interaction with the bridgehead enone and the enediyne moiety results in highly efficient DNA strand scission.2 Feeding 35S-labeled precursors to the culture, we were able to detect the incorporation of Na235S04, L- [3%]cysteine and L- [35S]methionine into esp A1 in the complex medium H946 (Table IV). The rates of incorporation of these three radiolabeled precursors were very low (0.019-0.095%). In an attempt to increase the rate of incorporation of 3%-precursors into esperamicins, the above ( 1 3) Williamson, J. M.; Inamine, E.; Wilson, K. E.; Douglas, A. W.; Liesch, J. M.; Albers-Schonberg, G. J . Biol. Chem. 1985, 260, 46374647. 1978, 100, (14) Weller, D. D.; Rinehart, K. L., Jr., J. Am. Chem. SOC. 6751-6160. . . . ..
(15) (a) Castle, M.; Blondin, G.; Nes, W. R. J . Am. Chem. SOC. 1963,85, 3306-3308. (b) Goad, L. J.; Hamman, A. S.A,; Dennis, A,; Goodwin, T. W. Nature 1966, 210, 1322-1324. (16) Lam, K. S.;Gustavson, D. G.; Forenza, S.Isolation of blocked mutants of esperamicin A1 from Actinomadura uerrucosospora. 46th annual meeting of the Society for Industrial Microbiology, Seattle, WA, August 1419.1989; Abstract P-84. (17) Rothstein, D. M.; Love, S.F. J . Bacteriol. 1991, 173, 7716-7718.
experiment was repeated by growing the culture in a defined medium. Defined medium DF-15, using sodium sulfate (0.2%) as the sole source of sulfur, supports the production of esp AI, but not esp Al in the fermentation.18 Adding Na235S04 into this defined medium yielded 1.7% incorporation of radioactivity into esp AI, (Table IV). ~ - [ ~ ~ S ] m e t h i o nand i n e~ - [ ~ ~ S ] c y s t ewere ine also incorporated into esp AI, at about the same efficiency as Na235S04 (Table IV). When L-methionine (0.2%)or L-cysteine (0.2%) was substituted for sodium sulfate as the sole sulfur source in the defined medium DF- 15, no esperamicins were detected even though the growth of the organism and the pH of the fermentation were the same as those in medium DF-15. The above data suggested that the sulfur from L-methionine or L-cysteine can only be incorporated into either the thiomethyl sugar or the trisulfide moiety but cannot provide all four sulfur atoms. Since sulfur from L-methionine has been shown to be efficiently incorporated into the thiosugar moieties in actinomycete fermentations,19 this may suggest that the allylic trisulfide cannot be derived from L-methionine or L-cysteine. An experiment was set up to test whether all four sulfur atoms of esperamicin can be derived from Na2S04. Na234S04was used to label esp Alcand the incorporation of the 34Sisotope evaluated by MS. Comparison of the MS fragments of the natural esp AI, (3%) with those of the labeled species (34S) allowed for rapid quantitation and location of the incorporated 34Satoms (Figure 4). [34S]Esp A], was prepared by growing the culture in defined medium DF-15 using Na234S04as the sole sulfur source. The full scan FAB mass spectrum of [j4S]espAI, showed a molecular weight of 1304 Da, 8 mass units higher than that for native esp AI,, suggesting the incorporation of four atoms of 34S. The presence of the four 34Satoms was confirmed by high resolution FAB-MS ([M H]+ m / z 1305.3708, calcd for C S ~ H T ~ N ~ O ~ ~ ~ ~ 1305.3743). The location of the labeled atoms was determined by comparison of the substructures of native and labeled compounds (Figure 4). The ions observed at m / z 1185 and 1187 result from the neutral loss of the allylic trisulfide from the protonated molecular ions of the native ( m / z 1297) and labeled ( m / z 1305) compounds, respectively. The 118 Da neutral loss of the allylic trisulfide from the labeled compound confirmed that all three sulfur atoms of the trisulfide were present as 34S atoms. The difference of 2 Da observed for the trisaccharide fragments confirmed the presence of 34Sin the thiosugar.
+
Conclusion We have demonstrated that the CIS bicyclic enediyne ring structure of esp Al is derived from head-to-tail condensation of seven acetate units and the uncoupled carbon is derived from C2 of acetate. The two carbons of the yne moieties of esp AI are derived from separate acetate units and are in good agreement with those of DNM-A.8 On the basis of the 13C-labeled acetate (18) Lam,K.S.;Veitch,J.A.;Golik,J.;Forenul,S.;Doyle,T.W.Production and isolation of two novel esperamicin analogs in a defined medium. 49th annual meeting of the Society for Industrial Microbiology, San Diego, CA, August 9-14, 1992; Abstract P-55. (19) Argoudelis, A. D.; Eble, T. E.; Fox, J. A.; Mason, D. J. Biochemistry 1969, 8, 3408-3415.
Lam
12344 J . Am. Chem. SOC., Vol. 115, No. 26, 1993
904+
OCHl
144
264
'
CH30
et
al.
obtained. The surface growth of the slant culture was transferred into a 500-mL Erlenmeyer flask, containing 100 mL of the vegetativemedium consistingof 2% starch, 0.5%glucose, 1% Pharmamedia, 1% yeast extract, and 0.2% CaC03. This vegetative culture was incubated at 28 "C for 96 h on a rotary shaker set at 250 rpm. Thevegetative culturewas mixed with an equal volume of cryoprotective solution consisting of 10%sucrose and 20% glycerol. Four-milliliter portions of this mixture were transferred to sterile cryogenic tubes (5" capacity) and were frozen in a dry iceacetone bath. The frozen vegetative cultures were stored at -80 "C until use. Media and Culture Conditions. A vegetative culture of strain MU5019 was prepared by transferring 4 mL of the cryopreserved culture to a 500-mL Erlenmeyer flask containing 100 mL of a vegetative medium. The vegetative culture was incubated at 28 "C and 250 rpm on a rotary shaker. After 96 h, 8-mL aliquots were transferred to 500-mL Erlenmeyer flasks containing 100 mL of esperamicin-producing medium. Two production media were used in this study. A complex medium, H946, was prepared using 6% cane molasses, 2% starch, 2% fish meal, 0.01% CuS04.5Hz0,0.2% CaCO3, and 0.00005% NaI. A defined medium, DF-15, consisted of 4% sucrose, 0.2% NHdC1,0.2% Na2S04, 0.1% K2-
HP0~,0.1%MgC1~,0.1%NaC1,0.2%CaC0~,0.0001%MnCl~,0.0001% -
Amrate
32 S Espemldn AI,
34S-Esperamicin A,,
Accurate
= 1296.3860, C57H,5N40&,
mas =
2.0p p
1304.3630, C S ~ H ~ N ~ O 2.6 ~ / ppm ~ S ~ ,
Figure 4. Summary of ions observed in the full-scan FAB mass spectra of [3zS]esperamicin AI, (MW 1296) and [34S]esperamicin AI, (MW 1304). The molecular ions of [34S] esperamicin AI, are in bold type.
enrichment pattern, we propose that the enediyne ring moiety of esp AI is derived from an octaketide with the loss of C1 of the end acetateunit as shown in Scheme I. Our data on the production of esp A I from cultures of A. uerrucosospora supplemented with cerulenin and sodium oleate rule out the possibility that the enediyne moiety of esp A1 is derived from the oleate+repenynate catabolic pathway. Since the labeling pattern of the two carbons of the yne moieties of NCS Chrom A7 is different from those of esp AI and DNM-A, this may suggest that the enediyne cores of esp AI and DNM-A are biosynthesized from a common precursor while N C S Chrom A is biosynthesized via a different process. The L- [methyl-l3C] methionine incorporation results show that the S-methyl groups of the trisulfide and the thiosugar groups and the 0-methyl groups of the aminosugar, the aromatic chromophore, and the carbamate moieties are derived from L-methionine via S-adenosylmethionine. Further work will be carried out to confirm the formation of the isopropyl group of the aminosugar from L-methionine. Understanding the mechanism of inhibition of esperamicin production by L-methionine may help us to design a way to relieve this inhibitory effect and control the formation of esp AI, esp Alb, and esp AI, in the fermentation. Initial studies carried out in our laboratory using Na234S04as the sole sulfur source in the fermentation have demonstrated that all four sulfur atoms in esp A,, can be derived from Na234S04. The above method requires only microgram quantities of 34Slabeled esperamicin. We thus hope to use a combination of 34Sprecursors, blocked mutants of sulfur metabolism of A. uerrucosospora, and FAB-MS to further study the biosynthesis of sulfur in esperamicin. Information concerningsulfur metabolism in actinomycetes is scarce, and it would therefore be valuable to explore the biosynthetic pathway leading from sulfate to the allylic trisulfide and the thiosugar in esperamicn from A. verrucosospora. Experimental Section Materials. Sodium salts of 90% enriched [1-'3C]-, [2-I3C]-,and [1,2I3C2]aceticacid were purchased from Merck Isotopes. N a ~ 3 ~ S 0was 4 obtained from Icon Company. ~-[3%]Methionine,~-[~sS]cysteine, and Na235SO4 were purchased from Du Pont NEN Research Products. Microorganism. A. uerrucosospora strain MU-5019 was used in this study. Strain MU-5019 wasmaintainedas a cryopreserved culturestored at -80 "C. To prepare a cryopreserved culture, strain MU-5019 was grown on slants for 2-3 weeks until uniform production of spores was
ZnC12, 0.0001% FeC12, and 0.00005% NaI. The production cultures were incubated at 28 "C and 250 rpm on a rotary shaker. Extraction and Analytical Method. The production of esp AI and esp AI, was monitored by HPLC using a C-18 reverse-phase column (Novapak, 3.9 X 150 mm2, Waters Associates). The solvent system was 0.05 M ammonium acetate (pH 4.5)/CH,OH/CH,CN (1:l:l) at a flow rate of 1 mL/min with the detector wavelength set at 254 nm. The fermentation extracts for HPLC assay were prepared by extracting the fermentation broth with an equal volume of ethyl acetate. The ethyl acetate extracts wereconcentrated 10-fold. Twenty-five to fifty microliters of the extracts were used for HPLC analysis. The amounts of esp Ai and esp At, in the extracts were determined by comparison with authentic standards. ['Vpcetate Incorporation. The time of addition of [I3C]acetate was carefully determined by monitoring the production of esp Ai during 2-5 days after inoculation of strain MU-5019 to the production medium. When theculture producedabout 0.14.2pg/mLespA1 (usually between 4-5 days), ["Clacetate was added to the culture to yield the final concentration of 0.2%. The production of esp Ai was then monitored daily. When the rate of production of esp Ai dropped to about 2-3 pg mL-l d a y i , 13C-labeledesp A1 was extracted from the culture. The titer of esp A1 at the time of harvest was about 18-20 pg/mL. L-[me~~~'~ethionineIncorporstion. The first portion ofL-[merhyl*3C]methioninewas added to the fermentation to yield a concentration of 0.025% when the production of esp A1 was about 0.1-0.2 pg/mL. The second portion of [iJCImethionine was added to the culture 24 h after the first addition to yield a final concentration of 0.05%. Purification of Biosynthetically l%-Labeled ESP A'. Culture broth (10 L) was extracted with an equal volume of ethyl acetate. After the extract was dried with anhydrous sodium sulfate, it was evaporated to dryness under reduced pressure. The dried extract was dissolved in 60 mL of CHC13 and mixed with 20 g of silica gel (LiChroprep Si 6 0 , 4 0 4 3 pm, Merck). The solvent was removed to absorb the esperamicins onto the silica gel. This was placed on top of a silica gel column comprising 100 g of silica gel in a 150-mL sintered glass funnel. The column was washed sequentially with 2 L of hexane, 3 L of toluene, 2 L of CCl,, 2 L of CH2C12, 3 L of CHClp, and 1 L of CHCl3/CH,OH (9:l). The CHC13/CH3OH (9:l) eluate, containing the esperamicin complex, was dried over anhydrous sodium sulfate and concentrated in uucuo. This residue was then dissolved in 30 mL of CHCl3 and mixed with 10 g of silica gel. The solvent was removed, and the silica gel containing the esperamicin complex was applied onto the top of a column of silica gel comprising 50 g of silica gel in a 60-mLsintered glass funnel. Thecolumn was successively developed using 1-L aliquots of increasing concentrations of acetone in hexane. Esperamicin complex was eluted in the 50%acetone fraction. The acetone fraction was treated with anhydrous sodium sulfate and concentrated in vacuo. The enriched espcramicin complex was further purified by reverse-phase chromatography using a Prep System Gold LC/system (Beckman) fitted with a Radial-Pak Cartridge (MBondapak CIS,8 X 100 mm2,Waters Associates). The solvent system was 0.05 M ammonium acetate/CHsOH/CH$N (3:2:5), and the detector wavelength was set at 320 nm. The flow rate was 10 mL/min. Esp AI was eluted at near 17 min. The fractions containing esp A1 were combined and concentrated in vacuo to remove CHJOH and CHJCN. The aqueous mixturewasextracted twicewith anequalvolumeof CH2C12. Theorganic
Biosynthesis of Esperamicin A , extract was then evaporated to dryness under reduced pressure to yield 25-30 mg of pure [I3C]esp AI. Purification of %-Labeled Esp At,. The production culture (100 mL) using medium DF-15 containing N a ~ ~ ~ was S 0 extracted 4 with 100 mLof ethyl acetate. The extract was evaporated to dryness under reduced pressure. Thedriedextract wasdissolved in 3 mL of CHCI3. Thesolution was applied to a column containing 10 g of silica gel (Lichroprep Si60, 40-63 pm, Merck) previously equilibrated with 100 mL of hexane. The column was washed sequentially with 200 mL of hexane, 100 mL of hexane/acetone (9: l), 40 mLof hexane/acetone (4:1), 40 mL ofhexane/ acetone (7:3), 40 mL of hexane/acetone (3:2), 40 mL of hexane/acetone ( l : l ) , and 40 mL of acetone. Esperamicins were recovered from the last three fractions, which were combined and evaporated to dryness. The extract was then dissolved in 0.5 mL of DMSO, and five aliquots of 0.1 mL each were injected into a C-18 reverse-phase column (Novapak, 3.9 X 300 mm2, 4 pm, Waters Associates) using the solvent system of 0.05 M ammonium acetate (pH 4.5)/CHsOH/CH3CN (1:l:l) at a flow rate of 2 mL/min. Fractions containing esp Alc were collected and dried under nitrogen. About 200 pg of esp AI, was recovered from HPLC. Conversionof Esp A1 to Macetyl-espAI. A solution of esp AI (30 mg) in dry methylenechloride (5 mL), cooled to 0 OC, was treated with acetic anhydride (100 pL) and 4-(dimethy1amino)pyridine (5 mg). Reaction was carried out for 16 h at 4 OC. The reaction was monitored by TLC on silica gel using 33% acetone in methylene chloride (esp AI, Rf= 0.25; diacetyl-espAI, Rf=0.74). After the reaction was completed, the mixture
J . Am. Chem. SOC.,Vol. 115, No. 26, 1993
12345
was filtered through a layer of silica gel (3 mL), which was then washed with methylene chloride. The products were eluted from the silica gel layer with 10 mL of acetone. The solution was concentrated in vacuo to a small volume and subjected to purification by preparative TLC (Kieselgel60 F254S, 20 X 20 X 0.05 cm3 plate, Merck) using 33% acetone in methylene chloride. The major product diacetyl-esp A1 (24 mg,80% yield) was eluted from the plate with acetone. The major byproduct triacetyl-esp A1 (5.3 mg,17.6% yield), with Rfof 0.84, was also isolated. WNMR Spectroscopy. The NMR experiments were performed either on a Bruker WM 360wb spectrometer or on a Bruker AM 500 spectrometer. IT-Enriched esp AI samples were dissolved in deuterated methanol under argon. Broad-band proton decoupled, ‘gated’, and onedimensional INADEQUATE 13Cspectra were obtained on the enriched samples using delays optimized for IJcc coupling of 83 or 166.7 Hz at ambient temperature. The carbon-carbon coupling constants were tabulated using the broad-band decoupled as well as the INADEQUATE spectra. All the chemical shifts are referenced to external TMS. Mass Spectrometry. Low- and high-resolution mass spectrometric analyses were performed on a Kratos MS50 mass spectrometer in the positive-ion mode using fast atom bombardment (FAB) ionization. The instrument was equipped with a saddle-field FAB gun (Ion Tech, Teddington, UK) operating at 8 KeV with xenon as the primary atom beam. m-Nitrobenzyl alcohol (NBA) was used as the matrix. Accurate mass measurements were obtained by peak matching with a cesium iodide saturated glycerol solution as the reference.
